7 Comments

Summary:

The Associated Press says it will use algorithm-generated content from Automated Insights to produce earnings reports, which has some professional journalists nervous about robots taking their jobs — but in reality such drudgery is better off being done by algorithms than by human beings

Ever since companies like Narrative Science started to become more well-known outside of data-science and machine-learning circles, journalists and fans of traditional media have expressed concern that robots — or rather, content-authoring software that is powered by algorithms — might someday take over the duties of real journalists. In fact, that’s been happening fairly steadily in certain journalism markets over the past several years, including sports journalism, and now there is another major entrant: Associated Press, the global newswire, has announced a content partnership with Automated Insights, an algorithm-driven content-production company.

In this case, as the Poynter Institute describes, the AP will be distributing stories that are automatically generated by Automated Insights (which Associated Press has acquired a small investment stake in) based on earnings reports from Zack’s, an investment research company. According to a post by AP on the announcement, the newswire service expects the new partnership will allow it to produce 10 times as many earnings reports as before — and managing editor Lou Ferrara said that soon a majority of AP earnings reports will be automated.

Although some journalists may be afraid that Automated Insights and Narrative Science will take away their jobs, Ferrara says the new partnership isn’t about cutting jobs, but about freeing up human reporters from having to engage in the drudgery of earnings reports, which in turn will allow them to devote their time to pieces that add value:

“Rather than spending a great deal of time focusing on the release of earnings and hammering out a quick story recapping each one, we are going to automate that process for all U.S. companies… our journalists will focus on reporting and writing stories about what the numbers mean and what gets said in earnings calls on the day of the release, identifying trends and finding exclusive stories we can publish at the time of the earnings reports.”

Narrative Science has been generating sports stories — game wrap-ups, etc. — and earnings-based reports for a number of clients including Forbes magazine for a couple of years now. In an interview with Steven Levy of Wired magazine, co-founder Kristian Hammond estimated that within 15 years robots or algorithms would be writing about 90 percent of what we consider news-based journalism. Is that a bad thing? It is if you endorse a kind of “full employment” principle for human journalists, but not if you actually care about journalism.

The harsh reality is that much of what appears in newspapers and on websites is not the kind of ground-breaking, investigative or analytical content most people think of when they hear the term “journalism.” Some of it is pedestrian content about sporting events, earnings reports, news releases, calendar events, city council meetings and so on. Wouldn’t it be better if we could automate some of that and free up reporters to do other things?

In a way, the debate about robot journalists is similar to the concern about the rise of “citizen journalism” or amateur journalism: namely, the fear among some traditional journalists that these new competitors will take jobs away from professionals. But by widening the pool of available reporters to include both amateurs and robots, we increase the amount of potential journalism being done — all it means is that as a professional journalist, you now have to make sure that you are better than a robot. If you aren’t, you should probably think about finding another line of work anyway.

Post and thumbnail images courtesy of Thinkstock / Ociacia

You’re subscribed! If you like, you can update your settings

Comment

Community guidelines
Thursday, August 28, 2014
you are commenting using your account. Sign out / Change

Comment using:

Or comment as a guest

Be sure to review our Community Guidelines. By continuing you are agreeing to our Terms of Service and Privacy Policy.

7 Comments

  1. Reblogged this on newarray and commented:
    This is a sad reality. No matter how people say it, robots will replace real journalists. That’s all scientists think about inventing… Robots.

  2. I guess this can work in sort of a just-the-facts/numbers way. No context, no explanation, no quotes. But then why not just create a chart with just the numbers and publish that? You can cram in even more earnings in fewer column inches that way. Chopping these numbers up into paragraphs is pointless without a good reason to do it, like a narrative story. Otherwise you’re wasting folks time. But I guess the goal is people want a story so let’s give them the best impression of one…instead of an actual one.

    1. exactly! a better approach would have been to survey readers to see how many do NOT read already-robotic-like earnings “reports” that offer nothing other than numbers in narrative form (ABC reported revs of [fill in the blank] which was [ditto] % higher than same period last year. Earnings were [ditto} etc etc…end of “report”).

      Like they yanked the daily stock prices, they can yank these too. Get some real writers to LISTEN TO THE EARNINGS CALLS and summarize those for us (i.e., tell us WHY revs and earnings moved in the direction they did)….

  3. SillyRabbit Tuesday, July 1, 2014

    This is something that Reuters has been doing for some time with SEC filings, and there are definitely pros and cons. The benefit is broader coverage about smaller companies, although if they didn’t merit coverage by humans before, very few people probably cared in the first place.

    The downside, which is significant, is the error rate that Reuters has had so far, and the relative difficulty getting those errors corrected. My understanding is that the algorithmic operation for them is in India, and the process for fixing mistakes has proven to be more time consuming for humans than anyone probably expected. So far, the robots are not nearly good enough.

    It seems like the AP is taking a bigger gamble on algorithmic reporting than even Reuters, which does say something about how much they value (or do not value) accuracy – the primary concern for readers of earnings stories.

  4. There are so few real journalist theses days that robots would be a huge improvement in the news department as well. Journalists cover up as many stories as they report on and are usually heavily biased. I would assume robots would do a much better job at day to day news and basic reporting, but there needs to be more honesty and non-biased journalism across the board.

  5. Richard Gaughan Wednesday, July 16, 2014

    The key here for journalists is that making “sure that you are better than a robot” may not provide value to the marketplace. Before becoming a science writer, I was an aerospace engineer, where the motto is “better is good enough.” The meaning being fairly clear: if you’ve got something that already meets your requirements then any resources you put into making it better is waste.

    Unfortunately, I found this has translated into the markets that had been my primary customers: the trade magazines. I used to write articles that would describe a technological or scientific innovation in enough detail for scientists and engineers to determine if the methods described could help them in their own work. Now the readers primarily read those magazines just to learn that something exists — and then they go to the internet for a more in-depth look (if the subject warrants). So the trade magazines now run what are essentially (and sometimes literally) press releases. Costs them nothing, and they are “good enough” to keep their subscribers renewing.

    Were my articles “better” than the press releases they now run? Perhaps not every one was a gem, but at least they had independent research and an objective perspective — traits that I believe the general reader would find “better.” But that “better” was more than “good enough,” so it is no longer saleable to those markets. Again, it’s unfortunate that “good enough” is acceptable in many more markets than those markets looking for “better.”

    I’m sure this pattern is the same in many other journalistic markets. And if a robot can be “good enough,” then “better than a robot” provides no value at all.

    http://www.mountainoptical.com/ScientificSeen/

    1. Oops! That quote from the aerospace industry is “better is the enemy of good enough.” Slightly different meaning without those key words!